Re: comedi as a POSIX standard

Baldwin, Stu W (JFTI) wrote:
> An IEEE Standard for Open Source software? Hey, where is that free beer,
> anyway? 

I don't see the irony. Big chunks of open source software already comply 
with certain key IEEE (and other) standards. For example:

http://www.fsf.org/manual/glibc-2.2.5/html_node/Standards-and-Portability.html#Standards%20and%20Portability

The rest of Stu's message seems (to me) to say I want a really good open 
specification, and I want it to be free. Great, me too. No one's 
stopping you. What I'm proposing is in no way contrary to that.

The issue boils down to objectives. If all we want is good documentation 
that all of us in the open source community can use, then all we need to 
do is help Herman get it written. In fact, we should help Herman in any 
case. Good documentation is the first step. Nothing will take away our 
own good documentation, and we can use it for whatever purposes we like, 
forever. Period.

I'm suggesting something more than that. I'd like to use what we have in 
Comedi to influence the commercial world, and in particular, to entice 
platform vendors (commercial RTOS, embedded systems, etc.) and a/d 
subsystems to implement and support the Comedi API. That'd be a good 
thing, in my opinion, even for those of us who have no plans to use 
Comedi outside open source. I write software for my users, and if they 
want to use my code on Solaris, who am I to say they shouldn't? If a 
large fraction of a/d boards came to market with vendor-provided comedi 
drivers, that'd be good too.

Now, if you don't think that's a good objective, then we disagree, and I 
propose to leave it at that.

If we agree on the objective, then the next issue is how to achieve it. 
I claim, and I believe there's evidence to support the claim, that 
commercial vendors are more likely to invest in the development required 
to support a formal standard from a recognized body than from Some Guys 
With A Web Site. In particular, Unix and RTOS vendors already have a 
demonstrated commitment to at least selectively complying with the 
POSIX.1 standard, which includes real-time functionality like clocks and 
timers, pthreads, etc. (For examples, Google for "wind river posix" or 
"lynx posix".) Commercial vendors support standards because they believe 
it leads to sales. The best way to get Comedi adopted seems to me to 
place it on the path they're already following.

The audience for a POSIX standard (in this scenario) is commercial 
vendors, and they don't mind paying for it. Ordinary users like us will 
continue to use the free documentation and be happy with it.

If I put an electrical outlet in my house, I don't need a copy of the 
National Electrical Code (which costs $$$) to comply with the code. The 
outlet vendor tells me how in their instruction sheet. That doesn't mean 
the NEC isn't important.

Steve

Received on 2003-04-14Z17:27:08